On October 19, 2026, the day of its opening keynote at IT Symposium/Xpo in Orlando, Gartner released results from its 2027 CIO and Technology Executive Agenda survey of 2,002 CIOs and technology executives, fielded from May through July 2026. The headline was blunt: only 13% of CIOs said their enterprise has realized significant or transformational value from AI so far.
Further down sits a finding that matters more for how you organize people. Organizations with explicit plans to redesign roles for a human-agent workforce are 2.4 times more likely to be what Gartner calls vanguards, the CIOs who deliver better business outcomes and whose CEOs see high value from AI. Only 4% of CIOs reported having such plans in place.
In the opening keynote the same day, Gartner cited a Q2 2026 survey of more than 2,000 CIOs and technology executives: 71% of organizations expect to deploy agentic AI within the next 12 months, but only 21% currently have agentic AI governance capabilities in place.
Put those numbers together, and the picture is clear. AI agents are about to join many engineering teams. The job descriptions, career ladders, and team structures those agents will work inside were written for a world without them.
A disclosure before we go further. Golabs provides dedicated AI teams and AI staffing, so we benefit when companies rethink how roles are designed and filled. Gartner sells research and advisory services, and McKinsey, which we also cite, sells organizational consulting. We flag each source’s interest where it applies.
What Does a Human-Agent Workforce Mean for an Engineering Org?
It means designing roles around the work that people and AI agents share, rather than adding agents to job descriptions written before the agents existed. In practice, that means fewer distinct role types, each with a wider span, plus clear human accountability for directing, checking, and correcting agent work. Gartner found only 4% of CIOs have explicit plans to do this.
Gartner describes the shift as roles that simultaneously compress and expand, with fewer distinct role types and a wider span of responsibility in each, while 40% of respondents still expect to increase IT headcount in 2027. In the Symposium keynote, Gartner described software developers being freed from the drag of manual coding to focus on rapid prototyping, requirements-driven development, architectural thinking, and product engineering.
McKinsey’s October 5, 2026 article on AI and organizational design gives one concrete example of what that looks like. In one organization’s product development function, traditional pods of eight to ten people shifted to hybrid teams of four to six people supported by agents.
That is a single example from McKinsey’s client work, not a survey finding, but it matches the direction we described in why AI teams are getting smaller and more senior. The point is not that teams shrink: each seat now carries more scope, and the person in it spends more time deciding, reviewing, and integrating than producing first drafts of code.
Why Do So Few Companies Have a Plan?
Because agents tend to arrive as tools, adopted team by team, rather than as an organizational decision. A team lead adds a coding agent, another team pilots an agent for test generation, and a third connects one to the ticketing system. Nobody rewrites a job description, because nobody decided to change the job.
The 71% versus 21% gap in Gartner’s keynote is the same pattern at the company level: deployment plans are running well ahead of the structures around them. We covered the control side of that gap in whether your controls kept up with multi-agent AI. Role design is the people side of the same problem.
McKinsey’s research points the same way. In its readiness panel survey, fielded from February to April 2026 among 750 English-speaking employees, 70% of respondents said they are personally ready for AI. Among a subsample of 608 leaders, only 27% said their organizations are ready to make the shifts required. Individuals are moving faster than the organizations they work for.
And the organizations getting results are the ones that changed the work itself. According to McKinsey’s 2026 State of AI survey, just 6% of respondents qualify as AI high performers, and nearly three-quarters of those high performers report substantially redesigning workflows, compared with only one-quarter of other organizations.
Which Engineering Roles Change First?
In our work staffing AI engineering teams, the changes show up in a predictable order. None of these roles disappears. Each one has a different center of gravity, and the job description, career ladder, and interview loop for each need to change accordingly.
Software engineers become owners of outcomes, not producers of code. When agents draft much of the code, the engineer’s value moves to specifying the problem well, judging the output, and integrating it safely. Career ladders that reward lines shipped or tickets closed start rewarding the wrong thing.
Reviewers and QA become evaluators. Checking agent output at volume requires test suites, evaluation sets, and clear acceptance criteria rather than line-by-line reading. That makes evaluation design a skill you hire and train for. We covered the testing side in how to test software when AI writes much of it.
Someone has to run the agents in production. Agents need monitoring, cost tracking, access reviews, and incident handling. At most companies we work with, that responsibility is currently not in anyone’s job description. It usually lands informally on whichever engineer set the agent up, which is exactly the accountability gap our agent governance guide warns about.
Coordination roles grow. McKinsey’s authors report that in their client work, coordinator roles have grown at roughly 1.5 to 2.0 times the rate of line roles over the past five years. When agents produce more work, more human time goes into deciding what gets built, by whom, and in what order.
Entry-level roles get redesigned, not deleted. If agents handle the tasks juniors used to learn on, the entry path has to be rebuilt around review, pairing, and supervised ownership. We made the case for keeping that pipeline in place should you still hire junior developers in the AI era.
Has Any Function Already Done This?
Customer service offers an early case, because AI reached its workload sooner than most functions. In a Gartner survey of 321 customer service and support leaders, fielded in September and October 2025 and released in April 2026, 85% said they are adding new tasks and responsibilities to frontline agent roles, and 75% are shifting agents into entirely new roles.
Only 31% have implemented or plan frontline workforce reductions via layoffs in response to AI by the first quarter of 2027, and 63% are reducing frontline headcount gradually through attrition rather than widespread cuts.
Kathy Ross, a Vice President Analyst at Gartner, put the warning plainly: service leaders need a plan for reshaping their workforce, “otherwise a plan will be handed to them.” That warning applies to engineering leaders too. If you do not redesign the roles, the redesign happens anyway, one tool purchase and one budget cut at a time.
How Do You Redesign Engineering Roles for 2027?
Map the work before the roles. List the work your engineering org does and mark where agents already touch it or will within a year: code generation, test writing, documentation, incident triage, data pipelines. Roles follow from that map. Starting from the current org chart tends to preserve it.
Give every agent a named human owner. For each agent in production, someone should be accountable for what it does, what it costs, and what happens when it fails. Write that accountability into a job description, with time allocated to it, rather than leaving it as an informal extra.
Consolidate role types on purpose. If Gartner is right that organizations will need fewer distinct role types, each with a wider span, decide which ones merge rather than letting it happen by attrition. For example, you might combine test automation and evaluation work into a single quality engineering role with a broader scope and a clearer bar.
Rewrite career ladders around judgment. Define levels by the scope of decisions a person owns, the quality of their review, and their ability to direct agent work, not by output volume. Promotion criteria are what determine whether a redesign becomes real or stays on a slide.
Rebuild the entry path. Pair early-career engineers with senior reviewers, give them supervised ownership of agent-assisted work, and make review skill something they are trained in. Without that path, today’s senior engineers have no one to hand their judgment to. Our AI fluency program is one structured option for building those skills in-house.
Decide which new seats to build, buy, or borrow. Some redesigned roles you will fill by training current staff, some by hiring, and some through partners during the transition. Our recent guide to planning AI skills for 2027 walks through that choice seat by seat, and a dedicated AI team can cover new seats while your own people grow into them.
Revisit the design every quarter. Agent capabilities change faster than annual planning cycles. Review which roles are working, which agents have owners, and where work is falling between seats, and adjust before next year’s budget locks the structure in place.
Common Questions About Redesigning Roles for a Human-Agent Workforce
What is a human-agent workforce?
A human-agent workforce is one where people and AI agents share the work, with roles designed around that split. People set direction, review output, and stay accountable for results, while agents carry out defined tasks. Gartner found that organizations with explicit plans to redesign roles this way are 2.4 times more likely to be AI vanguards.
How many companies have plans to redesign roles for AI agents?
Very few. Gartner’s 2027 CIO and Technology Executive Agenda survey of 2,002 CIOs and technology executives, fielded May to July 2026, found that only 4% of CIOs have explicit plans to redesign roles for a human-agent workforce. Yet 71% of organizations expect to deploy agentic AI within the next 12 months, according to a Q2 2026 Gartner survey.
Will AI agents reduce engineering headcount?
Not necessarily. In Gartner’s 2027 CIO Agenda survey, 40% of respondents expect to increase IT headcount in 2027, particularly in cybersecurity and data and analytics, while roles compress and expand. In our experience, the bigger change is in what each role does. Teams may get smaller per product while companies run more of them.
What new roles do AI agents create in engineering?
In our work, the new responsibilities center on running agents in production, evaluating agent output, and coordinating work among people and agents. Many companies fold these into existing roles rather than creating new titles. What matters is that each responsibility is written into a job description, with a named owner and time allocated.
If you are redesigning engineering roles for 2027 and want help deciding which seats to build, hire, or staff through a partner, talk to the Golabs team. We will start with the work map and tell you which seats should remain within your company.